Savin

@omriel1 we now have support for multi-cluster execution within a single flow - where different steps can run on different clusters. that would be one option to achieve your use...

@Sypek Are you using any Metaflow extensions? Would it be possible for you to share the output of `python flow.py step-functions create --only-json`?

Yep - would you like to contribute a PR?

lgtm! we can ship this once @saikonen approves

@HarryAnkers, do you have a reproducible example that demonstrates the speed-up? We don't set `shell=True` for many reasons (security, portability, etc.), and I am curious to dig into what is...

It seems that linter check failed. Can you address that so that we can merge this PR?

thanks for the PR! it might make more sense to introduce `tpu` as an arg rather than overriding `gpu` - since one can access gpus in gcp too. also, this...

@madhur-ob the message - `This could be a transient error. Use @retry to retry.` is incorrect if this is triggered through @timeout.

closes #1818 , #1608

Yep - this is the metadata service returning a 502. How is the service deployed right now?